systemsoftware
/
Programming-Guides
35
1
Fork 0
Code Issues 17 Pull Requests Projects Releases Wiki Activity

Change references to intrinsics into links

Browse Source 
Signed-off-by: Paul A. Clarke <pc@us.ibm.com>
pull/69/head
Paul Clarke 5 years ago committed by Bill Schmidt
parent 3f20e7da56
commit db465c1266
1 changed files with 96 additions and 82 deletions
Show Stats Download Patch File Download Diff File

178
Intrinsics_Reference/ch_biendian.xml
Unescape Escape View File

@ -180,8 +180,9 @@ vector unsigned __int128 x = { (((unsigned __int128)0x1020304050607080) &lt;&lt;
to access the <emphasis>N</emphasis>th vector element from a
vector pointer. The dereference operator <code>*</code> may
<emphasis>not</emphasis> be used to access data that is not
aligned at least to a quadword boundary. Built-in functions
such as <code>vec_xl</code> and <code>vec_xst</code> are
aligned at least to a quadword boundary. Built-in functions such as
<code><xref linkend="vec_xl" xrefstyle="select:title nopage"/></code> and
<code><xref linkend="vec_xst" xrefstyle="select:title nopage"/></code> and
provided for unaligned data access. Please refer to <xref
linkend="VIPR.biendian.unaligned" /> for an example.
</para>
@ -796,238 +797,238 @@ a[3] = c;</programlisting>
<tbody>
<row>
<entry>
<para>vec_bperm</para>
<para><code><xref linkend="vec_bperm" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_mergeh</para>
<para><code><xref linkend="vec_mergeh" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_signedo</para>
<para><code><xref linkend="vec_signedo" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_cipher_be</para>
<para><code><xref linkend="vec_cipher_be" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_mergel</para>
<para><code><xref linkend="vec_mergel" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_sld</para>
<para><code><xref linkend="vec_sld" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_cipherlast_be</para>
<para><code><xref linkend="vec_cipherlast_be" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_mergeo</para>
<para><code><xref linkend="vec_mergeo" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_sldw</para>
<para><code><xref linkend="vec_sldw" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_doublee</para>
<para><code><xref linkend="vec_doublee" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_mfvscr</para>
<para><code><xref linkend="vec_mfvscr" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_sll</para>
<para><code><xref linkend="vec_sll" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_doubleh</para>
<para><code><xref linkend="vec_doubleh" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_mule</para>
<para><code><xref linkend="vec_mule" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_slo</para>
<para><code><xref linkend="vec_slo" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_doublel</para>
<para><code><xref linkend="vec_doublel" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_mulo</para>
<para><code><xref linkend="vec_mulo" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_slv</para>
<para><code><xref linkend="vec_slv" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_doubleo</para>
<para><code><xref linkend="vec_doubleo" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_ncipher_be</para>
<para><code><xref linkend="vec_ncipher_be" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_splat</para>
<para><code><xref linkend="vec_splat" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_extract</para>
<para><code><xref linkend="vec_extract" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_ncipherlast_be</para>
<para><code><xref linkend="vec_ncipherlast_be" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_srl</para>
<para><code><xref linkend="vec_srl" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_extract_fp32_from_shorth</para>
<para><code><xref linkend="vec_extract_fp32_from_shorth" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_pack</para>
<para><code><xref linkend="vec_pack" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_sro</para>
<para><code><xref linkend="vec_sro" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_extract_fp32_from_shortl</para>
<para><code><xref linkend="vec_extract_fp32_from_shortl" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_pack_to_short_fp32</para>
<para><code><xref linkend="vec_pack_to_short_fp32" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_srv</para>
<para><code><xref linkend="vec_srv" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_extract4b</para>
<para><code><xref linkend="vec_extract4b" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_packpx</para>
<para><code><xref linkend="vec_packpx" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_sum2s</para>
<para><code><xref linkend="vec_sum2s" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_first_match_index</para>
<para><code><xref linkend="vec_first_match_index" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_packs</para>
<para><code><xref linkend="vec_packs" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_sums</para>
<para><code><xref linkend="vec_sums" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_first_match_or_eos_index</para>
<para><code><xref linkend="vec_first_match_or_eos_index" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_packsu</para>
<para><code><xref linkend="vec_packsu" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_unpackh</para>
<para><code><xref linkend="vec_unpackh" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_first_mismatch_index</para>
<para><code><xref linkend="vec_first_mismatch_index" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_perm</para>
<para><code><xref linkend="vec_perm" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_unpackl</para>
<para><code><xref linkend="vec_unpackl" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_first_mismatch_or_eos_index</para>
<para><code><xref linkend="vec_first_mismatch_or_eos_index" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_permxor</para>
<para><code><xref linkend="vec_permxor" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_unsigned2</para>
<para><code><xref linkend="vec_unsigned2" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_float2</para>
<para><code><xref linkend="vec_float2" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_pmsum_be</para>
<para><code><xref linkend="vec_pmsum_be" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_unsignede</para>
<para><code><xref linkend="vec_unsignede" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_floate</para>
<para><code><xref linkend="vec_floate" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_reve</para>
<para><code><xref linkend="vec_reve" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_unsignedo</para>
<para><code><xref linkend="vec_unsignedo" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_floato</para>
<para><code><xref linkend="vec_floato" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_sbox_be</para>
<para><code><xref linkend="vec_sbox_be" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_xl (ISA 2.07 only)</para>
<para><code><xref linkend="vec_xl" xrefstyle="select:title nopage"/></code> (ISA 2.07 only)</para>
</entry>
</row>
<row>
<entry>
<para>vec_gb</para>
<para><code><xref linkend="vec_gb" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_shasigma_be</para>
<para><code><xref linkend="vec_shasigma_be" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_xl_be</para>
<para><code><xref linkend="vec_xl_be" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_insert</para>
<para><code><xref linkend="vec_insert" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_signed2</para>
<para><code><xref linkend="vec_signed2" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_xst (ISA 2.07 only)</para>
<para><code><xref linkend="vec_xst" xrefstyle="select:title nopage"/></code> (ISA 2.07 only)</para>
</entry>
</row>
<row>
<entry>
<para>vec_insert4b</para>
<para><code><xref linkend="vec_insert4b" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_signede</para>
<para><code><xref linkend="vec_signede" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>vec_xst_be</para>
<para><code><xref linkend="vec_xst_be" xrefstyle="select:title nopage"/></code></para>
</entry>
</row>
<row>
<entry>
<para>vec_mergee</para>
<para><code><xref linkend="vec_mergee" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para></para>
@ -1056,7 +1057,8 @@ a[3] = c;</programlisting>
Before the bi-endian programming model was introduced, the
<code>vec_lvsl</code> and <code>vec_lvsr</code> intrinsics
were supported. These could be used in conjunction with
<code>vec_perm</code> and VMX load and store instructions for
<code><xref linkend="vec_perm" xrefstyle="select:title nopage"/></code>
and VMX load and store instructions for
unaligned access. The <code>vec_lvsl</code> and
<code>vec_lvsr</code> interfaces are deprecated in accordance
with the interfaces specified here. For compatibility, the
@ -1097,7 +1099,7 @@ a[3] = c;</programlisting>
<tbody>
<row>
<entry>
<para>vec_ld</para>
<para><code><xref linkend="vec_ld" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>lvx</para>
@ -1108,7 +1110,7 @@ a[3] = c;</programlisting>
</row>
<row>
<entry>
<para>vec_lde</para>
<para><code><xref linkend="vec_lde" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>lvebx, lvehx, lvewx</para>
@ -1119,7 +1121,7 @@ a[3] = c;</programlisting>
</row>
<row>
<entry>
<para>vec_ldl</para>
<para><code><xref linkend="vec_ldl" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>lvxl</para>
@ -1130,7 +1132,7 @@ a[3] = c;</programlisting>
</row>
<row>
<entry>
<para>vec_st</para>
<para><code><xref linkend="vec_st" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>stvx</para>
@ -1141,7 +1143,7 @@ a[3] = c;</programlisting>
</row>
<row>
<entry>
<para>vec_ste</para>
<para><code><xref linkend="vec_ste" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>stvebx, stvehx, stvewx</para>
@ -1152,7 +1154,7 @@ a[3] = c;</programlisting>
</row>
<row>
<entry>
<para>vec_stl</para>
<para><code><xref linkend="vec_stl" xrefstyle="select:title nopage"/></code></para>
</entry>
<entry>
<para>stvxl</para>
@ -1166,7 +1168,9 @@ a[3] = c;</programlisting>
</table>
<para>
Instead, it is recommended that programmers use the
<code>vec_xl</code> and <code>vec_xst</code> vector built-in
<code><xref linkend="vec_xl" xrefstyle="select:title nopage"/></code> and
<code><xref linkend="vec_xst" xrefstyle="select:title nopage"/></code>
vector built-in
functions to access unaligned data streams. See the
descriptions of these instructions in <xref
linkend="VIPR.vec-ref" /> for further description and
@ -1479,35 +1483,44 @@ a[3] = c;</programlisting>
<title>vec_sld and vec_sro are not bi-endian</title>
<para>
One oddity in the bi-endian vector programming model is that
<code>vec_sld</code> has big-endian semantics for code
<code><xref linkend="vec_sld" xrefstyle="select:title nopage"/></code>
has big-endian semantics for code
compiled for both big-endian and little-endian targets. That
is, any code that uses <code>vec_sld</code> without guarding
is, any code that uses
<code><xref linkend="vec_sld" xrefstyle="select:title nopage"/></code>
without guarding
it with a test on endianness is likely to be incorrect.
</para>
<para>
At the time that the bi-endian model was being developed, it
was discovered that existing code in several Linux packages
was using <code>vec_sld</code> in order to perform multiplies,
was using
<code><xref linkend="vec_sld" xrefstyle="select:title nopage"/></code>
in order to perform multiplies,
or to otherwise shift portions of base elements left. A
straightforward little-endian implementation of
<code>vec_sld</code> would concatenate the two input vectors
<code><xref linkend="vec_sld" xrefstyle="select:title nopage"/></code>
would concatenate the two input vectors
in reverse order and shift bytes to the right. This would
only give compatible results for <code>vector char</code>
types. Those using this intrinsic as a cheap multiply, or to
shift bytes within larger elements, would see different
results on little-endian versus big-endian with such an
implementation. Therefore it was decided that
<code>vec_sld</code> would not have a bi-endian
<code><xref linkend="vec_sld" xrefstyle="select:title nopage"/></code>
would not have a bi-endian
implementation.
</para>
<para>
<code>vec_sro</code> is not bi-endian for similar reasons.
<code><xref linkend="vec_sro" xrefstyle="select:title nopage"/></code>
is not bi-endian for similar reasons.
</para>
</section>
<section xml:id="VIPR.biendian.vperm">
<title>Limitations on bi-endianness of vec_perm</title>
<para>
The <code>vec_perm</code> intrinsic is bi-endian, provided
The <code><xref linkend="vec_perm" xrefstyle="select:title nopage"/></code>
intrinsic is bi-endian, provided
that it is used to reorder entire elements of the input
vectors.
</para>
@ -2533,7 +2546,8 @@ a[3] = c;</programlisting>
</tgroup>
</informaltable>
<para>
The lesson here is to only use <code>vec_perm</code> to
The lesson here is to only use
<code><xref linkend="vec_perm" xrefstyle="select:title nopage"/></code> to
reorder entire elements of a vector. If you must use vec_perm
for another purpose, your code must include a test for
endianness and separate algorithms for big- and

Write Preview
Loading…
Cancel
Save